Loïc Minier wrote: > I'm disturbed by this too, but -- as I clarified on IRC -- I think > there's a conflict of interests between getting more meaningful > backtraces in average (and hence improving the quality of Debian before > the release / saving ourself a message to bug submitters) and > testing/sid users as first class users and hence respecting policy > during the full release cycle. (My understanding is that the -dbg > recommend would have been dropped before the release.)
Note that you're making work for the installation team here. First, it's hard to tell which recommends are planned to be removed later, and this makes it hard to tell when there are few enough spurious recommends that d-i can turn on recommends by default without bloating the installation. Secondly, if we decided to ignore the gnome-dbg recommends since it will be removed, and we turned on recommends, we would be in a position where we couldn't properly test desktop installs for lenny stable, since the desktop installation would be much larger than it's planned to be in stable. So we'd not be able to test partman-auto recipes and so on to make sure they provide enough space; we'd not be able to catch whether tasksel disables the desktop task properly when there's not enough disk space; the installation manual wouldn't properly document how much space the desktop task uses; if we included recommends on CDs, we'd not be able to tell if gnome fit on a CD; and so on. Pushing all this work back to shortly before the next stable release is not a good thing. -- see shy jo
signature.asc
Description: Digital signature
